Anthropic won’t back down over Pentagon’s demands to remove AI safeguards

Share

Anthropic has refused to back down against a US government ultimatum to strip safety guardrails from its AI models.

Chief Executive Dario Amodei declared on Thursday that the company “cannot in good conscience” grant the US military unfettered access to its Claude AI, even as Defense Secretary Pete Hegseth threatens to terminate a $200 million contract and designate the firm a “supply chain risk.”

The conflict centres on two specific red lines drawn by Anthropic: the use of its technology for mass domestic surveillance of Americans and its integration into “fully autonomous” weapons systems capable of lethal action without human intervention.

Amodei argued that such applications are “simply outside the bounds of what today’s technology can safely and reliably do,” warning that autonomous weapons lack the critical judgment of trained troops and could put both warfighters and civilians at risk.

The Pentagon, however, views these safeguards as an impediment to national security, particularly as it races to keep pace with China’s AI developments. Defense officials, including Undersecretary Emil Michael, have criticized Anthropic for attempting to “personally control the US military,” asserting that the department should be trusted to follow existing laws.

The tension has been simmering for months, particularly after it was revealed that Claude was used in the classified operation to capture Venezuelan leader Nicolás Maduro.

The consequences of this defiance could be severe. Hegseth has given Anthropic until Friday evening to comply or face being labelled a supply chain risk – a designation usually reserved for foreign adversaries.

This would effectively blacklist Anthropic from doing business with any other vendor that works with the US military, potentially crippling its commercial prospects in the defence sector.

What happens next will serve as a definitive test of the AI industry’s independence. If Anthropic remains firm, the Pentagon may invoke the Defense Production Act to compel compliance or move to “offboard” the company entirely, shifting its classified operations to rivals such as Elon Musk’s xAI. 

For latest tech stories go to TechDigest.tv